Cloud Monitoring ListTimeSeries Basic Aggregations Reference

This document maps Cloud Monitoring Aligner and Reducer concepts to structured ListTimeSeries (list_timeseries) REST request fields. Use this matrix and default rules to determine the correct aggregation query parameters (aggregation.perSeriesAligner, aggregation.crossSeriesReducer, and aggregation.groupByFields) based on Cloud Monitoring metric properties (metricKind and valueType) and desired calculation goals.

Table of Contents

• Translation Matrix (~Line 26)
• Default Aggregations and Visualization Rules (~Line 74)
  • 1. CPU and Memory Utilization (Ratios / Percentages) (~Line 80)
  • 2. Rate of Events / Throughput (Counters) (~Line 99)
  • 3. Distribution Metrics (Quantiles / Latency) (~Line 114)
  • 4. Boolean & Status Metrics (BOOL Value Type) (~Line 132)
  • 5. Backlog Age & Processing Lag (ALIGN_MAX / REDUCE_MAX) (~Line 143)

Translation Matrix

To use this matrix:

• Inputs: Identify the metric's metricKind and valueType from its MetricDescriptor, then infer the target calculation goal from the user's prompt (e.g., Mean, Sum, 95th Percentile) under Aggregation Intent to select the matching perSeriesAligner and crossSeriesReducer.
• Output Aggregation Fields (MANDATORY): Extract BOTH perSeriesAligner AND crossSeriesReducer values from the table below to populate the aggregation query parameters (aggregation.perSeriesAligner and aggregation.crossSeriesReducer) in your ListTimeSeries REST request. Every request MUST specify both perSeriesAligner and crossSeriesReducer.

Default Aggregations and Visualization Rules

Apply these standard defaults when constructing ListTimeSeries REST query specifications for charts, dashboards, or when the user's aggregation preference is underspecified:

1. CPU and Memory Utilization (Ratios / Percentages)

• Use when: Querying CPU or memory utilization metrics (ratios or percentages) for any service or agent (e.g., compute.googleapis.com/instance/cpu/utilization or agent.googleapis.com/memory/percent_used).
• Default Aggregation Directive: For CPU and memory utilization metrics, default to perSeriesAligner = ALIGN_MEAN and crossSeriesReducer = REDUCE_NONE. If needed, group specifically by instance, such as groupByFields = ["resource.labels.instance_id"].
• Aggregation Constraints:
  • No Cross-Series Summing: Do NOT use crossSeriesReducer = REDUCE_SUM. Utilization metrics represent ratios or percentages; summing them across instances yields invalid percentages over 100%.
  • No Cross-Series Averaging for Resource Limits: Averaging utilization across instances (crossSeriesReducer = REDUCE_MEAN) masks severe outliers. For example, one instance crashing at 100% while others sit idle at 0%.

2. Rate of Events / Throughput (Counters)

• Use when: Querying DELTA or CUMULATIVE event counter metrics.
• Throughput Rule: You MUST convert DELTA or CUMULATIVE metrics representing event counts (INT64 / DOUBLE) to a rate by setting perSeriesAligner = ALIGN_RATE.
• Cross-Series Reducer: Use crossSeriesReducer = REDUCE_SUM when combining throughput across instances (e.g., total read bytes per second across all VMs in a zone).
  • Example: perSeriesAligner = ALIGN_RATE, crossSeriesReducer = REDUCE_SUM, alignmentPeriod = "300s".
• Removal of Transform Functions: Do NOT apply multi-layer transform aligners. Apply perSeriesAligner = ALIGN_RATE and crossSeriesReducer = REDUCE_SUM cleanly in a single primary aggregation query.

3. Distribution Metrics (Quantiles / Latency)

• Use when: Querying DISTRIBUTION metrics (such as request latencies).
• Rule: For DISTRIBUTION metrics, such as Cloud Run request latencies (run.googleapis.com/request_latency/e2e_latencies) or Pub/Sub ack latencies (pubsub.googleapis.com/subscription/ack_latencies):
  • To retrieve raw histogram bucket distributions across instances, use perSeriesAligner = ALIGN_DELTA (for DELTA/CUMULATIVE) or perSeriesAligner = ALIGN_SUM (for GAUGE) with crossSeriesReducer = REDUCE_SUM.
  • To extract specific percentile latency gauges directly via the API, use percentile aligners: perSeriesAligner = ALIGN_PERCENTILE_99, perSeriesAligner = ALIGN_PERCENTILE_95, perSeriesAligner = ALIGN_PERCENTILE_50, or perSeriesAligner = ALIGN_PERCENTILE_05. When reduction across instances is requested, combine with the matching reducer (e.g., perSeriesAligner = ALIGN_PERCENTILE_95, crossSeriesReducer = REDUCE_PERCENTILE_95).

4. Boolean & Status Metrics (BOOL Value Type)

• Use when: Querying BOOL value type metrics.
• Default Aggregations:
  • Fraction True / Availability: perSeriesAligner = ALIGN_FRACTION_TRUE, crossSeriesReducer = REDUCE_MEAN returns the fraction of healthy instances in [0.0, 1.0].
  • Count True: perSeriesAligner = ALIGN_FRACTION_TRUE, crossSeriesReducer = REDUCE_SUM returns the total count of healthy instances (INT64).

5. Backlog Age & Processing Lag (ALIGN_MAX / REDUCE_MAX)

• Use when: Querying metrics tracking maximum age, lag, or oldest unacknowledged items.
• Rule: For metrics tracking the maximum age, lag, or oldest unacknowledged item across services or workers (e.g., pubsub.googleapis.com/subscription/oldest_unacked_message_age or dataflow.googleapis.com/job/system_lag), always default to perSeriesAligner = ALIGN_MAX and crossSeriesReducer = REDUCE_MAX to surface peak delays across instances.